Problem
Research questions and friction points this paper is trying to address.
Ordinary Differential Equation Discovery
Pharmacological Digital Twins
Population Modeling
Sparse Noisy Data
APOD: reasoning-guided agentic population ordinary differential equation discovery for pharmacological digital twins
🤖 AI Summary
This study addresses the challenge that manually constructing population ordinary differential equation (ODE) models in pharmacology is time-consuming and often overlooks inter-individual variability. To overcome this, we propose an automated modeling framework driven by large language model (LLM) agents. This approach introduces a novel reasoning-guided agent search mechanism that integrates biological prior knowledge with Bayesian inference, extending beyond traditional constrained search spaces to automatically discover population digital twin ODE systems incorporating individual variation within an open space. Experimental results demonstrate structure recovery rates of 94%–100% and a twelvefold improvement in efficiency. Furthermore, the framework successfully predicts platelet toxicity and optimizes dosing regimens, providing an efficient new paradigm for precision pharmacology.
📝 Abstract
Establishing ordinary differential equations (ODEs) describing population data is a fundamental part of mathematical modeling in pharmacology, crucial to developing digital twins. However, doing so from sparse, noisy data is a slow, expert-driven task. Existing automated methods either search a restricted model space or ignore population inter-individual variability. Here we introduce APOD (Agentic Population ODE Discovery), a language-model agent that iteratively reasons over biological knowledge and fit diagnostics in an open-ended search space to discover a population digital twin (PDT), i.e., a shared ODE system with between-subject variability. On synthetic pharmacokinetic and tumor-dynamics benchmarks, APOD recovered ground-truth structures in 94-100\% of runs, 12-fold faster in median than an established library-based search. On real cohorts it converged to valid structures, and proposed a PDT of radioligand-therapy-induced platelet dynamics that predicts thrombocytopenia from first-cycle data and simulates alternative dosing schedules that lower the predicted risk of toxicity.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Ordinary Differential Equation Discovery
Large Language Model Agent
Population Digital Twin
Pharmacological Modeling
Inter-individual Variability
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
No related jobs found.
